* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
American Politics Research http://apr.sagepub.com Lobbying Strategies, Venue Selection, and Organized Interest Involvement at the U.S. Supreme Court Thomas G. Hansford American Politics Research 2004; 32; 170 DOI: 10.1177/1532673X03259192 The online version of this article can be found at: http://apr.sagepub.com/cgi/content/abstract/32/2/170 Published by: http://www.sagepublications.com Additional services and information for American Politics Research can be found at: Email Alerts: http://apr.sagepub.com/cgi/alerts Subscriptions: http://apr.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 10.1177/1532673X03259192 AMERICAN ARTICLEPOLITICS RESEARCH / MARCH 2004 Hansford / LOBBYING STRATEGIES LOBBYING STRATEGIES, VENUE SELECTION, AND ORGANIZED INTEREST INVOLVEMENT AT THE U.S. SUPREME COURT THOMAS G. HANSFORD University of South Carolina When and why will organized interests choose to lobby the U.S. Supreme Court by submitting amicus curiae briefs? This article argues that organized interests are most likely to lobby the Court when conditions increase their expectation of influencing the Court’s policy outputs. However, when deciding which policy venue to lobby, organized interests that rely on membership support will also have to consider the effect of their lobbying decisions on their ability to attract and retain members. Analysis of the amicus curiae brief filings of 579 organized interests suggests that an interest is more likely to submit amicus briefs at the Supreme Court when the Court is receptive to the positions advocated by the interest and the interest has participated at the Court in the past. The results also indicate that membership-based groups will take into account the extent to which relevant cases have been covered by the media when choosing whether to lobby the Court. Keywords: amicus curiae; Supreme Court; organized interests; lobbying strategies/tactics One the most important decisions that an organized interest makes involves selecting which policy venue, or set of venues, to lobby. Although traditional interest group research investigates group involvement in Congress or administrative agencies, the judicial branch also constitutes a policy venue in which organized interests can pursue their goals (Barker, 1967). Viewing the participation of organized interests in the judiciary as an important phenomenon, scholars have generated a substantial body of research indicating, for Author’s Note: An earlier version of this article was presented at the Conference on the Scientific Study of Judicial Politics, Columbus, Ohio, October 20-21, 2000. I wish to thank Dave Damore and Teena Gabrielson for their assistance with the data reliability analysis. I also appreciate the helpful comments provided by Lee Epstein, John Gates, Robert Jackman, Gary King, Brian Sala, Jim Spriggs, and the anonymous reviewers. AMERICAN POLITICS RESEARCH, Vol. 32 No. 2, March 2004 170-197 DOI: 10.1177/1532673X03259192 © 2004 Sage Publications 170 Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Hansford / LOBBYING STRATEGIES 171 instance, that a wide variety of organized interests participate at the Court as amicus curiae (Caldeira & Wright, 1990), that this involvement has increased over time (L. Epstein, 1993), and that organized interest advocacy may have some effect on Court decisions and opinions (L. Epstein & Kobylka, 1992; Kearney & Merrill, 2000; McGuire, 1990). It remains unclear, however, why organized interests choose to lobby the courts in the first place. In part, this question remains largely unanswered because most of the relevant scholarship focuses on detailing the participation histories of a fairly small set of liberal groups that frequently litigate (e.g., Ivers, 1995; Kobylka, 1991; O’Connor, 1980; Sorauf, 1976; Vose, 1959) instead of systematically examining the decisions made by all types of organized interests. Although the former approach provides useful information on the activities of an important and interesting subset of groups, it cannot answer the broader question of when and why the typical organized interest chooses to lobby the Court. Scheppele and Walker (1991) cast a wider net and utilized a survey of a large number of interest groups to test various hypotheses regarding why some groups indicate that litigation is important to their members. Their theory primarily focuses on considerations, such as the likelihood of meeting the requirement of standing, that are most relevant to the decision to use the courts by initiating lawsuits. I seek to build on the literature addressing the involvement of organized interests in the courts by examining the decision of organized interests to lobby the U.S. Supreme Court by submitting amicus curiae briefs. I begin by developing a general explanation of organized interest involvement at the Supreme Court. My theoretical starting point is that all organized interests seek policy influence, but some pursue this goal in a particularly constrained manner. Thus, one element of my theory focuses on the conditions under which organized interests can expect their lobbying activities to influence the Court’s policy outputs, whereas a second element, in contrast to prior research (e.g., Scheppele & Walker 1991), suggests that membership-based interests (as opposed to institutional interests) must take into account the effect of their lobbying activities on their ability to attract and retain membership support. Membership-based interests should be more likely to lobby the Court when this action allows the interest to Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 172 AMERICAN POLITICS RESEARCH / MARCH 2004 increase its provision of purposive incentives to members. By recognizing the differences between the lobbying decisions made by institutional interests and those made by membership-based groups, my theory applies to the broader organized interest population, not just a specific subset of it. I test my theoretical expectations with data on a large sample of organized interests and their amicus curiae brief filings at the U.S. Supreme Court from 1948 to 1995. These data allow me to avoid the selection bias problem encountered by research that focuses on a very limited set of fairly unrepresentative interest groups (such as the NAACP or the ACLU) that are known, ex ante, to be frequently involved in the courts. Even the more comprehensive studies typically examine only the organized interests that have participated in the courts (e.g., Caldeira & Wright, 1990; cf. Scheppele & Walker, 1991). My approach, in contrast, includes all types of organized interests, regardless of whether they are frequent participants at the Court. Thus, I am able to provide a particularly systematic examination of my hypotheses regarding the conditions under which organized interests decide to lobby the Court. ORGANIZED INTERESTS, VENUE SELECTION, AND THE SUPREME COURT When an organized interest participates in the policy process, it has to make a series of tactical decisions. This decision process begins with the organized interest choosing the policy venue, or set of venues, in which to focus its lobbying efforts. For example, the interest could opt to lobby Congress, the courts, a federal administrative agency, or some combination of these venues. The venue-selection decision is of critical importance for three reasons. First, this decision likely determines the extent to which the interest can attain its lobbying goals. Second, subsequent lobbying decisions made by the organized interest are shaped by the institutional characteristics of the venue in which it chose to become involved. Third, assuming that organized interest involvement can influence the policy process, the venue-selection decision has important ramifications for the policies set by American political institutions. Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Hansford / LOBBYING STRATEGIES 173 Whereas most research investigating the lobbying decisions made by organized interests focuses on legislative lobbying (e.g., AustenSmith & Wright, 1994; Hojnacki & Kimball, 1998), many interests also pursue their goals in the courts. Over the past few decades, organized interest involvement at the U.S. Supreme Court has risen dramatically with interests participating as litigants, litigant sponsors, and amicus curiae (L. Epstein, 1993). This last form of participation is the most frequent form of organized interest involvement at the U.S. Supreme Court, the kind of participation that most closely approximates traditional notions of lobbying activity, and the type of Supreme Court involvement typically studied by judicial scholars (e.g., Caldeira & Wright, 1988, 1990; L. Epstein & Knight, 1999; Kearney & Merrill, 2000; Krislov, 1963; Songer & Sheehan, 1993; Spriggs & Wahlbeck, 1997). Therefore, my analysis will focus on the filing of amicus curiae briefs.1 If organized interests have a choice regarding which policy venue(s) to lobby and the U.S. Supreme Court is one of the available venues, when will interests choose to lobby the Court by submitting amicus curiae briefs? To develop a theoretical answer to this question, I start with the assumption that organized interests pursue the goal of policy influence. An organized interest engages in lobbying activities to attain governmental policies that are as close as possible to the policy positions held by the interest. Policy influence is only likely to occur under certain conditions, so organized interests will prefer to lobby the policy venues in which these conditions are particularly favorable. As Salisbury (1984) noted, however, there is an important distinction to be drawn between organized interests of an institutional nature (e.g., corporations) and those that are membership-based groups (e.g., public interest groups). I argue that this distinction has implications for the lobbying strategies selected by organized interests. Specifically, organized interests that are membership-based groups have to consider the effect of their lobbying decisions on their ability to attract and retain membership support.2 Organized interests of an institutional nature do not have the same type of organizational maintenance concerns and, therefore, are able to behave in a less constrained manner when selecting which policy venues to lobby. In short, all interests base their venue-selection decisions on the extent to Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 174 AMERICAN POLITICS RESEARCH / MARCH 2004 which they can expect to influence policy outcomes. Membershipbased interests, however, must also keep an eye on the connection between lobbying activities and membership support. THE CONDITIONAL NATURE OF POLICY INFLUENCE At the Supreme Court, successful policy influence occurs when the information or argument provided by an organized interest has the intended effect on the Court’s holding or the legal rule established in the Court’s majority opinion.3 While seeking to fulfill their own policy goals, Supreme Court justices operate without complete information regarding the ultimate effect that a decision or legal rule will have. By submitting amicus curiae briefs, organized interests can provide the justices with information on the potential political, social, and legal implications of a Court ruling (Barker, 1967; L. Epstein & Knight, 1999). It is through the provision of this sort of information that organized interests can hope to influence the policy set by the Court. Indeed, Court opinions sometimes incorporate information provided by amicus curiae (Spriggs & Wahlbeck, 1997), and case studies indicate that the arguments made by organized interests can have a real effect on the legal rules adopted by the Court (e.g., L. Epstein & Kobylka, 1992). It is clear that at the Court, however, organized interests cannot always count on influencing policy. Instead, any influence resulting from the lobbying activities of organized interests will be quite conditional in nature. Under what conditions is policy influence most likely to occur? I argue that an organized interest is most likely to influence a venue’s policy outcome when the relevant policymakers are politically receptive to the positions advanced by the interest. The extent to which a policymaker is receptive to an organized interest’s lobbying efforts is determined by the degree of policy congruence between the policymaker and the interest as well as the reputation the interest has with the policymaker. To maximize policy influence, organized interests will seek out the policy venues in which these conditions are most favorable. An organized interest can expect to exert more influence when it lobbies policymakers who share the same policy positions as the inter- Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Hansford / LOBBYING STRATEGIES 175 est instead of policymakers who are indifferent or opposed to the interest’s positions. Generally speaking, a decision maker prefers to receive information from like-minded sources, because this type of information is most likely to prevent the decision maker from making a particularly suboptimal choice (Calvert, 1985). Spriggs and Wahlbeck (1997) provided tentative support for this argument by demonstrating that the Supreme Court’s majority opinions are more likely to incorporate the arguments made by amicus curiae that are ideologically similar to the Court. If organized interests seek to influence policy and this influence is most likely to occur with policymakers who share their policy positions, then interests will prefer to lobby sympathetic policymakers. Extant empirical research indicates that organized interests are more likely to lobby congressional committee members who are considered to be allies (Hojnacki & Kimball, 1998, 1999).4 I expect to see the same pattern when considering organized interests’ venue-selection decisions. That is, I expect organized interests to prefer to lobby policy venues that are sympathetic to the policy positions and agendas pursued by the interests. The extent to which a policy venue is congruent with an interest’s policy preferences will depend on two considerations: the policy (or ideological) position of the venue and the agenda pursued by the venue. The more proximate the policy position of the venue is to the position held by the interest, the more likely the interest is to lobby this venue. The more agenda space that the venue allocates to issues of concern to the organized interest, again, the more likely it is that the interest will choose to lobby this venue. Thus, when an organized interest decides whether to lobby the Supreme Court, it assesses the degree to which the Court has been ruling favorably in relevant cases and the proportion of the Court’s agenda that deals with issues of concern to the interest. Hypothesis 1a: The more congruent the Court’s policy position is with the position held by an organized interest, the greater the probability of the interest lobbying the Court by submitting amicus curiae briefs. Hypothesis 1b: The more congruent the Court’s agenda is with the agenda pursued by an organized interest, the more likely the interest is to lobby the Court. Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 176 AMERICAN POLITICS RESEARCH / MARCH 2004 It is important to note that the receptiveness of the Court along these two dimensions is only part of the overall picture. There are other available policy venues that an organized interest will need to evaluate to minimize the opportunity costs of lobbying the Court if there are other, more receptive venues. Even if the Court is moderately receptive to an organized interest’s policy positions, the interest may not lobby the Court if other policy venues are even more sympathetic. Alternatively, a moderately receptive Court might be the venue of choice for an organized interest when the other available venues are particularly unreceptive.5 Hypothesis 2a: The greater the congruence between an organized interest’s policy position and the position of an alternative policy venue (i.e., Congress or the presidency), the less likely it is that the interest will lobby the Court. Hypothesis 2b: The more congruent an alternative policy venue’s (i.e., Congress’s or the president’s) agenda is with the agenda pursued by an organized interest, the lower the probability of the interest lobbying the Court. The second factor that will affect whether an organized interest can expect to influence policy outcomes involves the prior relationship between the interest and the policymaker or policy venue. In general, the information provided by an organized interest is more likely to be incorporated by a policymaker if the interest and the policymaker have interacted repeatedly in the past. Information provided by repeat players is often considered to be more credible (Milbrath, 1963), in part because repeated messages involve a greater cost to the sender (Potters & Van Winden, 1992). If the justices are familiar with an organized interest because it has a history of lobbying the Court, then the information provided by the interest may be considered more reliable than the information presented by an interest with which the justices have no experience. Repeat players can expect to exert more influence over policy outcomes than single shotters (Galanter, 1974; cf., McGuire, 1995), and this suggests that prior involvement at the Court should increase the likelihood of future participation. Hypothesis 3: An organized interest will be more likely to lobby the Court in a given year if it has lobbied there in the past. Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Hansford / LOBBYING STRATEGIES 177 THE MEMBERSHIP CONSTRAINT A subset of the organized interest population will be particularly constrained when making lobbying decisions. Organized interests that are membership-based groups (as opposed to organized interests that are institutional in nature, e.g., corporations) will have to consider the effect of their lobbying activities on their ability to attract and retain membership support. There is fierce competition between membership groups over the finite set of potential members or supporters (Gates, 1998). If a group that relies on its members for necessary resources fails to maintain a certain minimal level of membership support, then it is likely to cease existing. At a minimum, a group struggling to sustain a membership base will have difficulty mounting effectual lobbying campaigns and, thus, will have a hard time achieving policy influence. Membership-based interests attract members by providing them with various selective incentives. In addition to material incentives, these groups often rely on purposive incentives to garner membership support (Clark & Wilson, 1961; Cook, 1984; Moe, 1980; Walker, 1991). A group provides purposive incentives by making its members feel that the group is actively pursuing its stated policy goals and is successfully influencing policy outcomes.6 In this sense, the provision of purposive incentives is tied to lobbying activities. If a membershipbased interest can engage in lobbying behavior that increases the amount of purposive incentives it can offer, it will have more success in attracting and retaining members. Therefore, when making choices regarding lobbying activities, a membership-based interest will need to take into account the impact of these choices on the provision of purposive incentives. There is not a perfect correlation between achieving policy influence and the provision of purposive incentives. If there was, membership-based interests could simply select lobbying activities based on the decision calculus discussed in the previous section. Instead, it is difficult even for policy experts to identify the effect of a group’s involvement on governmental policy. Current and potential members of an interest must find this even more problematic and clearly have incomplete information on the degree to which an interest is actually exerting any influence on policy outcomes. Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 178 AMERICAN POLITICS RESEARCH / MARCH 2004 How can a membership-based interest increase the amount of purposive incentives offered to its members? In the context of venue selection, the ability to provide purposive incentives will be related to the visibility of the policy venue lobbied and the policy debates occurring within the venue. Media coverage of relevant Court cases allows a membership-based interest to participate in visible policy conflicts. If a member is informed that the group fought for the member’s policy preferences or rights in Court cases with which the member is familiar because they were covered by the media, it will likely have some purposive meaning for the member. There will be fewer purposive benefits associated with claiming that the group is active in lobbying the Court if the member was previously unaware that relevant cases were being heard and decided by the Court and, thus, the member was unaware that the Court was a relevant battleground. A membershipbased interest will therefore see the Supreme Court as an attractive option when relevant Court cases have a high level of visibility and, to a large extent, this visibility depends on the level of media coverage the cases are given. Hypothesis 4: Membership-based interests are more likely to lobby the Court when relevant cases have been garnering media attention. DATA AND METHOD As previously mentioned, even the more comprehensive studies of organized interest involvement in the judiciary typically examine only the interests that did, in fact, participate in the courts (e.g., Caldeira & Wright, 1990).7 In an effort to avoid this type of selection bias, I randomly sampled 735 organized interests from Washington Representatives, regardless of whether they have been involved at the Court, and then compiled data on these interests and their participation histories at the Court.8 This sample includes a mix of membership-based interests (such as public interest groups and trade associations) and institutional interests (such as corporations and local governments). For 579 of these organized interests, I was able to gather data on whether they are membership-based or institutional, the year they were formed, the Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Hansford / LOBBYING STRATEGIES 179 issues with which they are concerned, and the positions taken on these issues.9 To generate several of the independent variables discussed below, it was necessary to develop a comprehensive issue typology that would accommodate the diverse set of interests represented in the sample. I accomplished this task by employing a slightly modified version of the issue typology used by Baumgartner and Jones in their Agendas Project data set.10 This modified typology encompasses 28 categories ranging from defense to crime to civil rights to agriculture.11 For each organized interest, I coded the issue(s) of concern to the interest.12 Guided by Spaeth’s (1998) definitions of liberal and conservative, I also coded the ideological direction (liberal, moderate, or conservative) of the position held by the interest on each of the relevant issues.13 For many of the organized interests in the sample, this meant content analyzing and categorizing the position statements provided by the interests. Many of the institutional interests, however, do not provide public statements regarding their policy interests and positions, so for these interests, I developed coding rules to impute the relevant issues for the interest and the positions that the interest holds. 14 Finally, it was also necessary to assign an overall ideological position for each organized interest. For this purpose, I used the following formula: [(# of issues in which the interest holds a liberal position) – (# of issues in which the interest holds a conservative position)]/ (# of issues with which the interest is concerned) With scores ranging from –1 to 1, any interest with a score less than –.15 was coded as holding a conservative position overall, any interest with a score between –.15 and .15 was coded as moderate, and any interest with a score greater than .15 was coded as liberal.15 DEPENDENT VARIABLE Using the United States Supreme Court Judicial Database—Phase II (Gibson, 1997), I obtained data on the amicus curiae briefs submitted on the merits for the 1953 through 1985 terms of the Supreme Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 180 AMERICAN POLITICS RESEARCH / MARCH 2004 Proportion of Interests Submitting a Brief .08 .06 .04 .02 0 1948 1955 1965 1975 Court Term 1985 1995 Figure 1: Proportion of Organized Interests Submitting Amicus Curiae Briefs, 1948 to 1995 NOTE: Each circle represents the proportion of organized interests that submitted at least one amicus curiae brief in the given Court term. The solid line connects running mean smoothed proportions. Court.16 I then extended these data by coding all the amicus curiae briefs filed on the merits from the 1948 to 1952 and 1986 to 1995 terms of the Court.17 By combining the amicus curiae data with the data on the sample of organized interests, I created a data set (N = 17,457) in which there is an observation for each interest for each year of its existence (bounded by 1948 and 1995). The dependent variable is whether an organized interest submitted an amicus curiae brief on the merits of a Supreme Court case in a given Court term. Of the 17,457 observations in the data, there are a total of 692 instances in which the dependent variable equals 1 (i.e., an organized interest submitted at least one amicus curiae brief to the Court in a given term).18 Figure 1 presents the proportion of organized interests submitting an amicus brief to the Court over the time span included in the data. This figure should be approached with some caution, because the proportions associated with earlier Court terms reflect the behavior of only a fraction of the organized interests included in the sample (i.e., the older organized interests).19 Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Hansford / LOBBYING STRATEGIES 181 INDEPENDENT VARIABLES Court policy congruence. This variable is measured as the percentage of relevant Court decisions at time t – 1 (i.e., the previous Court term) decided in a manner that is ideologically compatible with the organized interest.20 A liberal (conservative) decision is considered ideologically compatible for organized interests holding a liberal (conservative) position in the given issue area.21 If an interest holds a moderate position on an issue, then either a liberal or conservative decision is counted as being .5 of a compatible decision. A score of 0 indicates that the Court has ruled unfavorably in every single relevant case in the previous term, whereas a score of 100 indicates that all the relevant Court decisions in the previous term were favorable. I expect the coefficient for this variable to be positive in direction. Court agenda congruence. I measured this variable as the number of cases the Court heard at t – 1 within the issue areas relevant to the organized interest.22 This measure was then divided by the total number of issues of concern to the group.23 The result is a variable indicating the average number of cases heard by the Court per issue relevant to the interest at time t – 1. I expect there to be a positive relationship between this variable and the probability of an interest lobbying the Court. Congressional policy congruence. To create a variable for congressional receptiveness to an organized interest’s policy positions, I first identified the median score on the first DW-NOMINATE dimension (see Poole & Rosenthal, 1997) for each chamber of Congress for a given year. I then rescaled the scores to range from 0 to 100, conservative to liberal, and averaged the chamber medians to generate a measure of congressional ideology for the year. For a liberal organized interest, congressional policy congruence was determined by subtracting the most conservative observed value for Congress during the included time span (42.35) from this averaged median DWNOMINATE score at time t – 1. Thus, this variable equals 0 when Congress is at its most conservative (least receptive) position and equals 15.675 when Congress is at its most liberal (58.025 being the most liberal Congress observed in the data). For a conservative orga- Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 182 AMERICAN POLITICS RESEARCH / MARCH 2004 nized interest, this variable was determined by subtracting the ideological position of Congress from the most liberal observed value for Congress. For moderate organized interests, congressional policy congruence equals 15.675 minus the absolute deviation of the ideological position of Congress from 50 (the score of a perfectly moderate Congress). Thus, this measure is at its maximum (15.675) when Congress is moderate and will decrease in magnitude as Congress becomes particularly conservative or liberal. According to my argument, as congressional policy congruence increases in size, an interest should be less likely to lobby the Court. 24 Congressional agenda congruence. I measured the extent to which the congressional agenda is perceived to be receptive to an organized interest as the number of days of committee hearings at t – 1 that dealt with the issues of concern to the interest.25 This measure was then divided by the total number of issues relevant to the organized interest. I expect that an increase in congressional agenda congruence will decrease the probability of an interest lobbying the Court. Presidential policy congruence. This variable was generated in the same fashion as congressional policy congruence, except that here I use Poole and Rosenthal’s (1997) NOMINATE scores for the president.26 I assume that, although Congress can expand or contract the amount of discretion granted an agency, the preferences of the president determine, to a large extent, the policy positions and agendas of the executive branch (see D. Epstein & O’Halloran, 1996; Wood & Waterman, 1991). Because this variable increases in size as the president’s policy position becomes more congruent with an interest’s position, I expect there to be a negative relationship between this variable and the likelihood of an organized interest lobbying the Court. Presidential agenda congruence. According to Light (1999), the presidential state of the union address is “the statement of [a president’s] legislative priorities” (p. 160). Therefore, I content analyzed the presidential state of the union addresses to generate a measure of the presidential agenda. For each state of the union address, I coded the issue content of each sentence and assigned it 1 of the 28 issue categories in the modified Baumgartner and Jones typology.27 Presiden- Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Hansford / LOBBYING STRATEGIES 183 tial agenda congruence is measured as the percentage of sentences at t – 1 dealing with the issue(s) relevant to the organized interest in question. My theoretical framework suggests that the estimate for this variable should be negative. Previous Court involvement. This variable consists of the average number of amicus briefs filed at the Supreme Court by the interest over the prior 3 years. Put differently, this variable is measured as a 3year moving average of the number of amicus briefs filed per term by the interest (at times t – 3, t – 2, and t – 1). Previous Court involvement should have a positive coefficient, because I expect prior participation to increase an organized interest’s expectations of influencing the Court’s policy outcomes. Membership-based interest. This variable is coded as 1 if the organized interest has voluntary members and 0 otherwise. Thus, a group like Citizens for Law and Order is coded as 1, whereas the Continental Grain Company is coded as 0.28 Media coverage. I employed the data on The New York Times coverage of Supreme Court cases provided by L. Epstein and Segal (2000) to generate a measure of issue-specific media coverage of Court cases. More specifically, this variable equals the proportion of Court cases at time t – 1 falling in the issue areas relevant to an organized interest that were covered on the front page of the The New York Times. If relevant cases in the previous term received media coverage, the Court represents a more attractive venue for a membership-based interest concerned with participating in visible policy conflicts. Therefore, I interact media coverage with membership-based interest and expect this interaction term to have a positive coefficient. MODEL SPECIFICS To test my hypotheses, I estimated a time-series cross-sectional logit model in which the dependent variable is whether an organized interest submitted an amicus curiae brief on the merits of a Supreme Court case in a given Court term.29 In addition to the independent variables suggested by my hypotheses, the model also includes the com- Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 184 AMERICAN POLITICS RESEARCH / MARCH 2004 ponents of the interaction term (membership-based interest and media coverage) and a control variable for whether the organized interest is a state government (state government). This latter control is included because states participate at the Court at a much higher rate than other types of organized interest. To allow the baseline probability of participation at the Court to vary over time, I included calendar time in the model as a quadratic function in which calendar time is a countervariable starting at 1 in the year 1948 and calendar time squared is simply the square of this variable.30 The nature of time-series cross-sectional analysis raises a potentially serious problem of serially correlated errors, which can bias estimates of the standard errors and can lead to inefficiency in the parameter estimates (Beck, Katz, & Tucker, 1998). To ameliorate the problem of serial correlation with a binary dependent variable, Beck, Katz, and Tucker (1998) suggested the inclusion of a spline function as an independent variable that captures the effects of temporal dependence. Therefore, I generated a variable denoting the number of years since entry into the data set or since the organized interest last submitted an amicus curiae brief (duration). I initially included this variable in the model as a spline function. Surprisingly, the estimates for this spline function indicated the existence of a quadratic relationship leading me to ultimately include duration and duration squared as opposed to the less parsimonious spline segments. As an additional measure, I also estimate robust standard errors. RESULTS Estimates from the logit model explaining the likelihood of an interest filing a brief are displayed in Table 1. The statistically significant chi-squared statistic (p ≤ .05) indicates that the model has more explanatory power than a constant-only model. More importantly, the estimated coefficients for many of the independent variables of interest conform to my theoretical expectations. The results of the model support my contention that organized interests will respond to changes in the political receptiveness of the Court. The coefficient for Court policy congruence is positive and statistically significant, suggesting that the more congruent the Court’s Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Hansford / LOBBYING STRATEGIES 185 TABLE 1 Logit Model of the Probability of an Organized Interest Submitting an Amicus Curiae Brief at the U.S. Supreme Court, 1948-1995 Independent Variable Court policy congruence Court agenda congruence Congressional policy congruence Congressional agenda congruence Presidential policy congruence Presidential agenda congruence Previous Court involvement Media Coverage × Membership-Based Interest Component terms: Media coverage of Supreme Court cases Membership-based interest Constant Number of observations Log likelihood Chi squared (Wald test, 15 df) Estimated Coefficient Robust Standard Error 0.006† 0.022† –0.005 –0.000 –0.005 –0.004 2.13† 1.43† 0.003 0.009 0.015 0.001 0.003 0.013 0.256 0.711 0.443 0.168 –5.68† 0.457 0.195 0.427 17,457 –1627 1230† NOTE: The model also includes the following control variables (estimate, standard error): Calendar time (0.135, 0.027), Calendar time squared (–0.002, 0.000), State government (2.89, 0.245), Duration (–0.134, 0.017), and Duration squared (0.002, 0.000). †p ⱕ .05, one-tailed test. policy position is on issues relevant to an organized interest, the more likely it is the organized interest will choose to lobby the Court. The estimate for Court agenda congruence indicates that the degree to which the Court’s agenda deals with the issues of importance to an interest also exerts a positive effect on the probability of the interest submitting an amicus curiae brief.31 The results for the variables tapping the receptiveness of alternative policy venues are not as supportive of my argument. Although all four of the coefficients are in the predicted direction, none of the estimates reach a conventional level of statistical significance. The ideological position and agenda of the president and Congress appear to have little effect on the decision to submit amicus curiae briefs at the Court. These findings may, however, result in part from the necessarily rough measures employed.32 I have argued that the political receptiveness of the Court to an organized interest will also depend on the interest’s prior lobbying Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 186 AMERICAN POLITICS RESEARCH / MARCH 2004 activities, as there are greater potential benefits to lobbying a venue that has been lobbied in the past. Consistent with this hypothesis, the coefficient for previous Court involvement is positive and statistically significant. Organized interests exhibit a greater probability of filing amicus curiae briefs at the Court if they have lobbied this venue in the past.33 I have also contended that organized interests that are membershipbased will have to consider the impact of their lobbying activities on their ability to attract and retain members. My results provide evidence for this argument. The estimate for the interaction between membership-based interest and media coverage is positive and significant and thus supports my hypothesis that membership-based interests will respond to the amount of media coverage given to relevant Court decisions. As I expected, media coverage does not appear to influence the lobbying decisions made by institutional interests, as the estimate for the component term media coverage is not statistically significant. Although the direction and statistical significance (or insignificance) of the independent variables in the model provides useful information, it is also important to consider the substantive effect of these variables. Table 2 presents discrete changes in the predicted probability of a membership-based organized interest submitting an amicus curiae brief at the Court. These discrete changes in probability are calculated in the following manner. For each statistically significant independent variable of interest (except previous Court involvement, which I will discuss separately), I compute the probability of lobbying the Court for four different values of the variable: its minimum value, one standard deviation below its mean, one standard deviation above its mean, and its maximum value.34 Probabilities are presented for membership-based interests, because this allows the effect of the media coverage interaction term to be illustrated. From the predicted probabilities listed in Table 2, it appears that the media coverage interaction term has the largest substantive effect size. Although the discrete changes in probability presented here seem quite small, they increase substantially as the baseline probability increases. The substantive effect of any one of these variables is much greater when the values of the other independent variables combine to raise the baseline probability of an interest lobbying the Court.35 To Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Hansford / LOBBYING STRATEGIES 187 TABLE 2 Predicted Probabilities of a Membership-Based Interest Submitting an Amicus Curiae Brief Independent Variable Probability of Submitting a Brief Court policy congruence Minimum (0) – σ (31) + σ (60) Maximum (100) .088 .105 .123 .153 Court agenda congruence Minimum (0) – σ (.1) + σ (9) Maximum (48) .103 .104 .124 .257 Media Coverage × Membership-Based Interest Minimum (0) – σ (—) + σ (.2) Maximum (1) .098 .— .135 .416 NOTE: For each of the independent variables, probabilities are presented for four different values of the variable: its minimum value, one standard deviation below its mean, one standard deviation above its mean, and its maximum value. For the interaction term, subtracting a standard deviation from the mean value yields a value below the minimum of the range. See text for further discussion. illustrate, if a membership-based interest has filed an amicus brief at the Court in each of the 3 preceding years (prior Court involvement = 1), then the probability of lobbying the Court when relevant cases have not been covered in the media (media coverage = 0) is .240. If media coverage is set to its maximum observed value (1), then this probability increases to .659. It is also noteworthy that, although the probabilities presented in Table 2 are for a membership-based interest, the substantive impact of the variables (except for media coverage) is analogous for institutional interests. I now turn to the effect of previous Court involvement. I hypothesized that the prior lobbying history of an organized interest will influence its subsequent lobbying decisions, and the substantive effect of this variable is quite sizable. If an organized interest has filed an average of two amicus briefs per year in the 3 preceding years (previous Court involvement = 2) and other independent variables are set at their Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 188 AMERICAN POLITICS RESEARCH / MARCH 2004 means as above, there is a .746 probability of submitting an amicus curiae brief in the next year. This probability is much greater than the .042 probability associated with an interest that has not lobbied the Court in the past 3 years. Duration and duration squared, which are included to control for temporal dependence, act to condition the impact of previous Court involvement. In fact, the estimates for the duration variables indicate a curious type of temporal dependence as the probability of an interest lobbying the Court temporarily decreases as the length of time since the interest last submitted a brief (or entered the data set) increases. However, after a certain point (duration equals 28 years) this trend actually reverses, and the organized interest gradually becomes more likely to lobby the Court. Time also plays a role in the model through the inclusion of the control variables for calendar time and calendar time squared. It has been argued that, over the last few decades, organized interests have increasingly come to view the Court as a legitimate, useful lobbying target (L. Epstein, 1993). The estimates for calendar time and calendar time squared indicate that, once the other independent variables are controlled for, the likelihood of an organized interest lobbying the Court increased from 1948 to 1988, after which the probability of participation actually began to decrease. DISCUSSION When will an organized interest decide to lobby the U.S. Supreme Court by submitting amicus curiae briefs? My time-series crosssectional analysis indicates that an organized interest is more likely to participate at the Court as amicus curiae when the Court is politically receptive and when the interest has repeatedly interacted with the Court. Thus, it does appear that organized interests consider the likelihood of policy influence when making their venue-selection decisions. There is a running debate in the interest group literature regarding whether organized interests prefer to target policymakers that are sympathetic, indifferent, or hostile (see Austen-Smith & Wright, 1994, 1996; Baumgartner & Leech, 1996; Hojnacki & Kimball, 1998). Although the results of my analysis do not directly speak to this debate, they do suggest that organized interests are more likely to Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Hansford / LOBBYING STRATEGIES 189 lobby an institution populated with sympathetic policymakers. These results also intimate that, although relatively static considerations, such as the intensity of the policy conflict in which an interest group is involved, may make courts institutionally compatible in one sense (Scheppele & Walker, 1991), institutional compatibility is also dynamic. As there is turnover on the Court and the preferences of the justices change, the congruence between an interest’s policy position and agenda and those of the Court changes accordingly. There is less solid, empirical support for my argument that organized interests will consider the political receptiveness of alternative policy venues when deciding whether to file amicus briefs at the Court. Thus, my analysis may provide an interesting modification to the traditional political disadvantage theory, which emphasizes the political receptiveness of the other branches of government (Cortner, 1968). Although according to this theory groups go to the Court when the other branches of government are perceived as unfavorable, my analysis suggests that organized interests go to the Court when the Court is particularly favorable. My measures of the receptiveness of the other branches of government are somewhat blunt, however, and this may contribute to the statistical insignificance of the estimates for some of these variables. The results of my analysis provide support for my contention that membership-based interests will behave differently than their institutional counterparts. Indeed, membership-based interests are more likely to lobby the Court when the media are covering relevant cases. This aspect of my theory and analysis stands in contrast to most prior research on the involvement of organized interests at the Court, which has not recognized the differences in lobbying behavior that may result from the membership constraint. The principal implication here is that membership groups cannot pursue policy influence in an optimal manner. Instead, unlike corporations and other types of institutional interests, membership-based interests’ lobbying activities are constrained by the need to maintain membership support. There are other implications to be drawn from this analysis. For instance, Walker (1991) argued that interest groups select a lobbying strategy early in their existence and then stick with this strategy. My results can be interpreted as providing some support for this contention, as interests appear significantly more likely to lobby the Court if Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 190 AMERICAN POLITICS RESEARCH / MARCH 2004 they have lobbied there in the past. However, there is also a substantial dynamic element to this decision, because all organized interests respond to shifts in the receptiveness of the Court and a significant subset of membership-based interests react to changes in the way the media are covering Court cases. Finally, the results of my analysis imply that, when considering the effect of organized interest involvement on Court outcomes, it is important to be aware that the interests lobbying the Court in a given year do not represent a random sample of the organized interest universe. Instead, there is a systematic selection effect present. For example, the organized interests that submit amicus curiae briefs to the Court are likely to be somewhat ideologically closer to the Court than the average interest. If the justices are paying attention to the information provided in amicus curiae briefs, they are getting information from a less than representative sample of organized interests. NOTES 1. Although I am fundamentally interested in explaining the decision to participate at the Court as amicus curiae, Scheppele and Walker (1991) found that “overwhelmingly, those who submit amicus briefs also use the courts more directly [i.e., participate as litigants]” (p. 178). Therefore, my theory and empirical results may also, albeit somewhat indirectly, speak to the general decision of an interest to become involved in judicial advocacy more broadly defined. 2. When using the term membership, I do not intend to invoke the specific legal meaning that this term connotes for nonprofit corporations. Political scientists have used the term membership for the purposes of differentiating between interests that are membership based and those that are institutional (see Salisbury, 1984). Some scholars characterize organizational maintenance as an important goal for organized interests (Caldeira & Wright, 1995; Clark & Wilson, 1961; Gates, 1998). There is empirical evidence, however, that group leaders prefer to use available resources to seek policy influence (e.g., Lowry, 1997). Surveys of interest group leaders also indicate that they are primarily interested in influencing policy outcomes (e.g., Sabatier & McLaughlin, 1990). Thus, I approach organizational maintenance as an important constraint, not a goal. Organizational maintenance is not an end but a means to an end. 3. This statement refers to the nature of policy influence at the merits and opinion-writing stages of the Court’s decision-making process. Organized interests can also influence Court decision making by affecting the Court’s agenda-setting decisions (Caldeira & Wright, 1988). 4. There is not complete consensus on this issue. Notably, Austen-Smith and Wright (1994) argued that organized interests prefer to lobby unsympathetic legislators and will only lobby sympathetic legislators when they are being lobbied by opposing interests. 5. This element of my argument is similar to the political disadvantage theory first advanced by Cortner (1968). According to Cortner, organized interests will turn to the courts Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Hansford / LOBBYING STRATEGIES 191 when other policy venues are unsympathetic to their claims. Although the political disadvantage theory focuses on the receptiveness of alternative venues, I expect the receptiveness of the Court to be at least as important. 6. Moe (1980) argued that, because of incomplete information and the strategic behavior of group leaders, rational individuals may have an exaggerated sense of the effect of their involvement on the group’s ability to achieve its policy goals. 7. Scheppele and Walker (1991) adopted a somewhat mixed approach. They included both judicially involved and noninvolved interest groups when modeling the importance of litigation to groups. When examining the relative importance of amicus curiae participation, Scheppele and Walker chose to include only the groups that reported using the courts. 8. There are three main listings of organized interests that have been used in prior research: Washington Representatives, the Washington Information Directory, and the Encyclopedia of Associations. For my purposes, Washington Representatives is preferable, because it is the only one that includes institutional interests. 9. I primarily relied on four sources of information. First, I collected data from a variety of publications such as the Encyclopedia of Associations and Public Interest Law Groups: Institutional Profiles. Second, I consulted the materials published by the organized interests themselves (brochures, Web pages, etc.). Third, I used information provided by federal agencies. This included, for example, information from the Securities and Exchange Commission regarding public companies. Finally, I was also able to contact directly many of the interests and have them provide information that I could not obtain elsewhere. 10. This data set is available at http://depts.washington.edu/ampol/agendasproject.html. 11. I took the 19 major topic codes outlined by Baumgartner and Jones and subdivided several of them. The resulting 28 categories include macroeconomics, civil rights/minority issues, civil liberties, health, agriculture, labor/employment, immigration, education, environment, energy, transportation, law/crime, family issues, social welfare, community development/housing, banking/finance, domestic commerce, defense, space/science technology, communication, foreign trade, international affairs, general government operations, federalism, separation of powers, regulation of campaigns/elections, public lands/water management, and Indian affairs. Summary statistics indicate that civil rights/minority issues, civil liberties, regulation of campaigns and elections, and Indian affairs are the issue areas in which the highest proportion of organized interests in my data file briefs. 12. To assess the reliability of my coding of the relevant issues for an interest, I had a second coder code the issues for a random sample of 25 of the interests. The rate of agreement between my coding of the issues and that of the reliability coder ranges from 88% to 100%. To determine the degree to which these rates of agreement compare with the agreements expected by chance, I calculated kappa statistics (see Cohen, 1960) for each issue category. For all but two (space/ science/technology and public lands/water/natural resources) of the 28 issue areas, kappa is greater than .6 indicating substantial agreement between the two coders (Landis & Koch, 1977). Thus, I conclude that my assignment of issue categories to the organized interests in my sample is sufficiently reliable. It is possible that my use of contemporary issue positions will lead to some measurement error for observations occurring earlier in the 1948 to 1995 time span, because current issue priorities for an interest may not reflect past issue priorities. However, I think it reasonable to assume that the great majority of organized interests do not change the fundamental issues that drive their advocacy activities. The 28 issue categories I use are sufficiently broad to allow some movement regarding the exact issues with which a group is concerned without causing an unreliable classification. Nonetheless, I conducted a test to see if this measurement strategy appears to affect my results. I generated a dummy variable that equals 1 if the year in which the observation occurs is Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 192 AMERICAN POLITICS RESEARCH / MARCH 2004 greater than or equal to 1985. I then interacted this dummy variable with all of the independent variables that involve the issue codes (e.g., Court policy congruence, Court agenda congruence, etc.) and estimated my logit model with these interaction terms as well as all of the original independent variables. If the use of contemporary issue priorities to project past issue priorities leads to substantial measurement error, then the addition of these interaction terms should improve model fit, because the effects of the independent variables should be conditional on whether the observation in question occurs in a year relatively close to the year in which I coded the issue position statements (i.e., whether the observation occurred between 1985 and 1995 as opposed to prior to 1985). A likelihood ratio test reveals that model fit is not significantly improved (p > .05) by including these interaction terms. I repeated this test two more times, first with a dummy variable dividing the data at 1980 and then with a dummy variable dividing the data at 1975. For these other specifications, the interaction terms also yield no improvement in model fit. I take this as at least tentative evidence that variables relying on the issue codes do not perform better for recent years or underperform in more distant years. 13. These issue position codes were also subjected to a reliability analysis. A second coder coded the policy positions for the same 25 randomly selected interests. There are four possible codes for each issue area for each interest: liberal position, moderate position, conservative position, and issue not relevant to the interest. Intercoder agreement again ranges from 88% to 100%, and all but two of the kappa statistics (those for space/science/technology and public lands/ water/natural resources) are greater than or equal to .6. Based on these results, it appears that these data are reliable. 14. This imputation process was fairly straightforward. For example, all businesses were coded as having interest in the general economic policy categories (such as domestic economic policy and macroeconomics) as well as any specific categories into which the business might fall (e.g., agriculture, energy, transportation, etc.). When imputing the policy positions held by institutional interests, I used a second set of coding rules. For example, businesses were coded as holding conservative positions on economic issues (see McCarty & Poole, 1999; Schlozman & Tierney, 1986), and cities were coded based on the policy positions specifically outlined by the National League of Cities. 15. I chose to employ three discrete categories as opposed to a continuous measure of interest ideology, because the continuous measure is too sensitive to the number of issues relevant to the organized interest. The cut points between the three categories are necessarily somewhat arbitrary, but they were selected because the distribution of the continuous score is clearly trimodal and the cut points represent fairly obvious dividing lines between the modes. When generating the ideological position of an organized interest, I weight the positions held by the interest equally across all issues of concern to the interest. Although organized interests concerned with multiple issues may weight these issues differently, I do not think that this assumption is particularly problematic for my measure of the ideological position of organized interests. Because I ultimately make trichotomous the ideological position of organized interests, differences in the weighting of issues within one of the three categories will not matter. That is, if an interest is interested in issues X, Y, and Z and holds conservative positions on all three, it will not matter how the importance of these issues is weighted. Of the organized interests in my data, 232 (40.1%) hold issue positions that are all in the same ideological direction. For these interests, the equal weighting assumption cannot be problematic. Of the interests, 514 (88.8%) have consistent preferences across at least 75% of the issues that are relevant to them. That is, liberal interests hold liberal positions for at least 75% of the issues that matter to them. Thus, it is unlikely that weighting issue importance differentially would change the ultimate classification of the vast majority of interests in my sample. Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Hansford / LOBBYING STRATEGIES 193 16. Amicus curiae briefs are sometimes filed by coalitions of organized interests. In this situation, Gibson’s (1997) data included the first 11 signers of the brief. For briefs that had more than 11 signers, I coded all additional signers from Lexis and Briefs and Records of the United States Supreme Court. When coding the dependent variable, all signers of the brief are considered as having submitted the brief. 17. These data were gathered from United States Reports, Lexis, and Briefs and Records of the United States Supreme Court. I do not include any amicus curiae briefs filed at the certiorari stage of the Court’s decision-making process. 18. With rare events data, logit coefficient bias can be exacerbated (King & Zeng, 2001). For this reason, I also generated estimates using King and Zeng’s (2001) rare events logit model and compared these estimates with those obtained with the traditional logit model. There is very little difference between the two sets of estimates, which suggests that the distribution of the dependent variable is not leading to any significant bias. As a result, I present the traditional logit estimates here. It is possible that some interests may never file a brief at the Court. But just because an interest has not submitted a brief during a period of time does not mean that it was not at risk for submitting a brief during that time or in the future. The probability of an interest filing an amicus brief may be very small, but surely very few interests actually have a zero probability of filing a brief. Even if there are interests in my sample for which the probability of filing a brief in a given year is always zero, then the inclusion of these interests in my analysis should act to deflate the coefficient estimates for my independent variables. This means that my significance tests should be quite conservative, which is not undesirable. Further, the inclusion of the previous Court involvement variable (essentially a lagged dependent variable of sorts) in my model should help to minimize the heterogeneity that would be introduced if there were, in fact, interests for which the risk of filing a brief is always zero. 19. The median date of origination for organized interests in my sample is 1966. 20. If, for a given year, the Court did not hear a single case in the issue areas relevant to a particular interest, I used the Court’s overall percentage of liberal decisions across all issue areas to approximate policy congruence for that year. All the substantive variables are lagged (t – 1) for two related reasons. First, when making a decision at time t, it is not necessarily safe to assume that the interest will have information on the Court’s decisions, for example, that are being made at the same time (or actually in the near future). Second, by lagging these variables, it is more reasonable to assume that the independent variables are causing the dependent variable and not the other way around. 21. I used Spaeth’s (1998, 1999) coding of the ideological direction of a Court decision. 22. I determined whether a case dealt with an issue relevant to an organized interest by matching Spaeth’s (1998, 1999) issue codes to the modified Baumgartner and Jones issue typology. 23. To control for variation in the number of issues relevant to the organized interests in the sample, all of the measures of agenda congruence are divided by the number of issue areas of concern to the interest. 24. McCarty and Poole (1999) found that PAC contributions to members of Congress can be explained with a single, dimensional model utilizing NOMINATE scores. Adding a second dimension does not add explanatory power. This suggests that it is acceptable to place a variety of organized interests and members of Congress on one dimension. Because the NOMINATE scores for presidents are in the same metric and on the same dimension as the scores for Congress, McCarty and Poole’s results also suggest that it should be acceptable to place organized interests and the president on the same single dimension. Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 194 AMERICAN POLITICS RESEARCH / MARCH 2004 25. The data on congressional committee hearings were originally collected by Frank R. Baumgartner and Bryan D. Jones with the support of National Science Foundation grant number SBR 9320922 and were distributed through the Center for American Politics and Public Policy at the University of Washington. 26. NOMINATE scores for presidents only go back as far as Eisenhower. I generated a NOMINATE score for Truman by regressing the NOMINATE scores on Zupan’s (1992) measure of presidential ideology (which includes Truman). I then use this regression model to predict a NOMINATE score for Truman. If I estimate my model while excluding the Truman years, the substantive results for presidential policy congruence do not change. 27. Some sentences in each address could not be placed in the issue typology and were coded as “no issue content.” To assess the reliability of this measure, I had a second coder code 75 randomly selected sentences from the 1946 to 1995 presidential state of the union addresses. For each sentence, 1 of 29 codes could be assigned (28 issue codes and a code for no issue content). The rate of agreement between the codes assigned by the coder and the codes I assigned was 93.3%. This yields a kappa statistic of .93, which indicates nearly perfect agreement (Landis & Koch, 1977). Therefore, I conclude that these data are highly reliable. 28. Because membership in labor unions is often not completely voluntary, unions are coded as 0. Organized interests that do not technically have members but actively seek contributions are coded as 1, because, functionally speaking, they will behave like a membership group and compete for voluntarily contributed resources. 29. There is a second possible way in which my data could be approached. I could attempt to develop a selection model in which the first stage predicts whether the interest will be somehow involved in the courts and the second stage predicts the filing of amicus curiae briefs at the Supreme Court. I believe my single-stage model approach is sufficiently appropriate, however. A two-stage model would make more sense if I were ultimately interested in explaining why organized interests file amicus curiae briefs instead of initiating lawsuits or sponsoring litigants, because the dependent variable in the second stage of the model would distinguish between different forms of judicial advocacy. 30. Prior research implies that the propensity of organized interests to participate at the Court has increased over time (L. Epstein, 1993). There is no reason to expect that this propensity has increased in a linear manner, and experimentation with other functional forms suggests that a quadratic formulation is appropriate. 31. It is plausible that the effect of policy congruence will be conditioned on agenda congruence and vice versa. I therefore generated three interaction terms (Court Policy Congruence × Court Agenda Congruence, Congressional Policy Congruence × Congressional Agenda Congruence, and Presidential Policy Congruence × Presidential Agenda Congruence) and estimated my full model with these three variables also included. None of the estimated coefficients for the interaction terms are statistically significant, and a likelihood ratio test indicates that the inclusion of these three variables does not improve the fit of the model. Because I have no theoretically compelling reason to include these interaction terms and there is no empirical support for their inclusion, I only report the results of the additive specification. 32. The insignificant estimate for the two presidential variables may also be a function of the possibility that some organized interests (i.e., those interested in regulatory policies) are more likely to lobby the executive branch regardless of the extent to which this branch is receptive to the interest’s positions. The coefficient estimates for these two variables may be attenuated by the inclusion of interests that will primarily lobby the executive branch regardless of its receptiveness to their claims. I should also note that, whereas the results for all of the other independent variables are quite robust across various model specifications, the results for congressional policy congruence are Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Hansford / LOBBYING STRATEGIES 195 sensitive to model specification. Under certain specifications, the coefficient for this variable is statistically significant. 33. Although I argue that prior Court involvement increases the probability of filing a brief because repeat players can expect to have a greater chance of influencing the Court’s policy outputs, it is also possible that prior involvement increases the likelihood of subsequent involvement because prior involvement makes future involvement less costly. That is, interests may develop infrastructure or contacts geared toward litigation, and this could reduce the start-up costs associated with filing amicus curiae briefs. Unfortunately, I cannot disentangle these two explanations. Further, it is difficult to determine whether previous values of a dependent variable cause subsequent values of the dependent variable or whether the residuals are simply correlated across time. I theoretically expect that the former scenario applies here and have taken care to reduce the potential problems associated with serially correlated errors, but there is no way to formally separate true state dependence from spurious state dependence. For these reasons, the result regarding previous court involvement should be taken as providing only tentative evidence for my hypothesis. 34. Previous Court involvement is set at .5, duration is set at 1, state government is set at 0, and all other independent variables are set at their means. 35. The discrete changes in probability shift as the baseline probability changes, because in a logit model, the relationship between the independent variables and the dependent variable is fundamentally nonlinear. Independent variables will have their greatest effect when the baseline probability of an event occurring is .5 and will have the smallest effect when the baseline probability approaches 0 or 1. REFERENCES Austen-Smith, D., & Wright, J. R. (1994). Counteractive lobbying. American Journal of Political Science, 38, 25-44. Austen-Smith, D., & Wright, J. R. (1996). Theory and evidence for counteractive lobbying. American Journal of Political Science, 40, 543-564. Barker, L. J. (1967). Third parties in litigation: A systematic view of the judicial function. Journal of Politics, 29, 41-69. Baumgartner, F. R., & Leech, B. L. (1996). The multiple ambiguities of “counteractive lobbying.” American Journal of Political Science, 40, 521-542. Beck, N., Katz, J. N., & Tucker, R. (1998). Taking time seriously: Time-series—cross-section analysis with a binary dependent variable. American Journal of Political Science, 42, 12601288. Caldeira, G. A., & Wright, J. R. (1988). Organized interests and agenda setting in the U.S. Supreme Court. American Political Science Review, 82, 1109-1127. Caldeira, G. A., & Wright, J. R. (1990). Amici curiae before the Supreme Court: Who participates, when, and how much? Journal of Politics, 52, 782-806. Caldeira, G. A., & Wright, J. R. (1995). Lobbying for justice: The rise of organized conflict in the politics of federal judgeships. In L. Epstein (Ed.), Contemplating courts (pp. 44-71). Washington, DC: CQ Press. Calvert, R. L. (1985). The value of biased information: A rational choice model of political advice. Journal of Politics, 47, 530-555. Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. 196 AMERICAN POLITICS RESEARCH / MARCH 2004 Clark, P. B., & Wilson, J. Q. (1961). Incentive systems: A theory of organizations. Administrative Science Quarterly, 6, 129-166. Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and Psychological Measurement, 20, 37-46. Cook, C. E. (1984). Participation in public interest groups: Membership motivation. American Politics Quarterly, 12, 409-430. Cortner, R. C. (1968). Strategies and tactics of litigants in constitutional cases. Journal of Public Law, 17, 287-307. Epstein, D., & O’Halloran, S. (1996). Divided government and the design of administrative procedures: A formal model and empirical test. Journal of Politics, 58, 373-397. Epstein, L. (1993). Interest group litigation during the Rehnquist Court era. Journal of Law and Politics, 9, 639-717. Epstein, L., & Knight, J. (1999). Mapping out the strategic terrain: The informational role of amici curiae. In C. W. Clayton & H. Gillman (Eds.), Supreme Court decision-making: New institutional approaches (pp. 215-235). Chicago: University of Chicago Press. Epstein, L., & Kobylka, J. F. (1992). The Supreme Court and legal change: Abortion and the death penalty. Chapel Hill: University of North Carolina Press. Epstein, L., & Segal, J. A. (2000). Measuring issue salience. American Journal of Political Science, 44, 66-83. Galanter, M. (1974). Why the “haves” come out ahead: Speculations on the limits of legal change. Law and Society, 9, 95-160. Gates, J. B. (1998). The motivations for interest group litigation in a changing interest group environment. Unpublished manuscript, University of California, Davis. Gibson, J. L. (1997). United States Supreme Court judicial database—Phase II. Ann Arbor, MI: Inter-University Consortium for Political & Social Research. Hojnacki, M., & Kimball, D. C. (1998). Organized interests and the decision of whom to lobby in Congress. American Political Science Review, 92, 775-790. Hojnacki, M., & Kimball, D. C. (1999). The who and how of organizations’ lobbying strategies. Journal of Politics, 61, 999-1024. Ivers, G. (1995). To build a wall: American Jews and the separation of church and state. Charlottesville: University Press of Virginia. Kearney, J. D., & Merrill, T. W. (2000). The influence of amicus curiae briefs on the Supreme Court. University of Pennsylvania Law Review, 148, 743-855. King, G., & Zeng, L. (2001). Logistic regression in rare events data. Political Analysis, 9, 137163. Kobylka, J. F. (1991). The politics of obscenity: Group litigation in a time of legal change. New York: Greenwood Press. Krislov, S. (1963). The amicus curiae brief: From friendship to advocacy. Yale Law Journal, 72, 694-721. Landis, R. J., & Koch, G. G. (1977). The measurement of observer agreement for categorical data. Biometrics, 33, 159-174. Light, P. C. (1999). The president’s agenda: Domestic policy choice from Kennedy to Clinton (3rd ed.). Baltimore: Johns Hopkins University Press. Lowry, R. C. (1997). The private production of public goods: Organizational maintenance, managers’ objectives, and collective goals. American Political Science Review, 91, 308-323. McCarty, N. M., & Poole, K. T. (1999). An empirical spatial model of congressional campaigns. Political Analysis, 7, 1-30. McGuire, K. T. (1990). Obscenity, libertarian values, and decision making in the Supreme Court. American Politics Quarterly, 18, 47-67. Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution. Hansford / LOBBYING STRATEGIES 197 McGuire, K. T. (1995). Repeat players in the Supreme Court: The role of experienced lawyers in litigation success. Journal of Politics, 57, 187-196. Milbrath, L. W. (1963). The Washington lobbyists. Chicago: Rand McNally & Co. Moe, T. M. (1980). The organization of interests: Incentives and the internal dynamics of political interest groups. Chicago: University of Chicago Press. O’Connor, K. (1980). Women’s organizations’ use of the courts. Lexington, MA: Lexington Books. Poole, K. T., & Rosenthal, H. (1997). Congress: A political-economic history of roll call voting. New York: Oxford University Press. Potters, J., & Van Winden, F. (1992). Lobbying and asymmetric information. Public Choice, 74, 269-292. Sabatier, P. A., & McLaughlin, S. M. (1990). Belief congruence between interest-group leaders and members: An empirical analysis of three theories and a suggested synthesis. Journal of Politics, 52, 914-935. Salisbury, R. H. (1984). Interest representation: The dominance of institutions. American Political Science Review, 78, 64-76. Scheppele, K. L., & Walker, J. L., Jr. (1991). The litigation strategies of interest groups. In J. L. Walker, Jr. (Ed.), Mobilizing interest groups in America (pp. 157-183). Ann Arbor: University of Michigan Press. Schlozman, K. L., & Tierney, J. T. (1986). Organized interests and American democracy. New York: Harper & Row. Songer, D. R., & Sheehan, R. S. (1993). Interest group success in the courts: Amicus participation in the Supreme Court. Political Research Quarterly, 46, 339-354. Sorauf, F. J. (1976). The wall of separation: The constitutional politics of church and state. Princeton, NJ: Princeton University Press. Spaeth, H. J. (1998). United States Supreme Court judicial database, 1953-1996 terms (8th ed.). Ann Arbor, MI: Inter-University Consortium for Political & Social Research. Spaeth, H. J. (1999). Expanded United States Supreme Court judicial database, 1946-1968 terms (4th ed.). Ann Arbor, MI: Inter-University Consortium for Political & Social Research. Spriggs, J. F., Jr., & Wahlbeck, P. J. (1997). Amicus curiae and the role of information at the Supreme Court. Political Research Quarterly, 50, 365-386. Vose, C. E. (1959). Caucasians only: The Supreme Court, the NAACP, and the restrictive covenant cases. Berkeley: University of California Press. Walker, J. L., Jr. (1991). Mobilizing interest groups in America. Ann Arbor: University of Michigan Press. Wood, B. D., & Waterman, R. W. (1991). The dynamics of political control of the bureaucracy. American Political Science Review, 85, 801-828. Zupan, M. A. (1992). Measuring the ideological preferences of U.S. presidents: A proposed (extremely simple) method. Public Choice, 73, 351-361. Thomas G. Hansford is an assistant professor of political science at the University of South Carolina. His research interests include the involvement of organized interests at the U.S. Supreme Court, the interpretation of precedent at the Court, and the potential for the other branches of government to shape and constrain the judiciary. His work has been published in the Journal of Politics, Law and Society Review, and Political Research Quarterly. Downloaded from http://apr.sagepub.com at Univ of California Merced on May 28, 2007 © 2004 SAGE Publications. All rights reserved. Not for commercial use or unauthorized distribution.